Plot method for check DAGs — plot.see_check

The plot() method for the performance::check_dag() function.

# S3 method for class 'see_check_dag'
plot(
  x,
  size_point = 20,
  size_text = 4.5,
  colors = NULL,
  which = "all",
  effect = "total",
  check_colliders = TRUE,
  ...
)

Arguments

x: A check_dag object.
size_point: Numeric value specifying size of point geoms.
size_text: Numeric value specifying size of text elements.
colors: Character vector of length five, indicating the colors (in hex-format) for different types of variables, which are assigned in following order: outcome, exposure, adjusted, unadjusted, and collider.
which: Character string indicating which plot to show. Can be either "all", "current" or "required".
effect: Character string indicating which effect for the required model is to be estimated. Can be either "total" or "direct".
check_colliders: Logical indicating whether to highlight colliders. Set to FALSE if the algorithm to detect colliders is very slow.
...: Currently not used.

Value

A ggplot2-object.

Examples

library(performance)
# incorrect adjustment
dag <- check_dag(
  y ~ x + b + c,
  x ~ b,
  outcome = "y",
  exposure = "x"
)
dag
#> # Check for correct adjustment sets
#> - Outcome: y
#> - Exposure: x
#> 
#> Identification of direct and total effects
#> 
#> Incorrectly adjusted!
#> To estimate the direct and total effect, at least adjust for `b`. Currently, the model does not adjust for any variables.
#> 
plot(dag)


# plot only model with required adjustments
plot(dag, which = "required")


# collider-bias?
dag <- check_dag(
  y ~ x + c + d,
  x ~ c + d,
  b ~ x,
  b ~ y,
  outcome = "y",
  exposure = "x",
  adjusted = "c"
)
plot(dag)


# longer labels, automatic detection of outcome and exposure
dag <- check_dag(
  QoL ~ age + education + gender,
  age ~ education
)
plot(dag)